Search Done Right

Enterprise search – or search within businesses – is a decade behind web search in terms of usage. Interesting questions are why? and whither? – what’s the trend?

There is roughly one web search per person per day in the U.S., counting the web searches at Google, Yahoo, Live, etc., but excluding searches at eBay, YouTube, CMU.edu, WashPost.com, USA.gov and the like. The analogue in business is searches done on the general intranet search, not at point solutions like Outlook search, desktop search, single-repository search, and so on. Our experience is that daily, general intranet searches lag web searches by orders of magnitude.

For example, a case study on Organon (a pharma company now part of Schering Plough) reveals a daily rate of twelve searches per day before their search overhaul, for a company that has over 10,000 employees. If you think this is uncommon, ask your friends who work at large companies what they think of their intranet search.

Why has enterprise search lagged so far behind web search?

Some assert that enterprise search is harder than web search, but I disagree. Web search solves hard problems: huge scale, overcoming spam and even porn, and delivering results across the globe in sub-second times. Enterprise search confronts great variability in corporate environments so that, absent a well-architected search platform which can flexibly handle the variability, deployment times are very long and completed implementations become frozen since no one can understand the system. And of course enterprise search must deal with secure content.

Tellingly, a Wall Street Journal article of a couple of years ago cited a VP of Marketing, responsible for his company’s intranet search, who explained that secure enterprise content was purposely NOT indexed by their enterprise search because of a lack of confidence in the engine’s ability to deliver content only to those authorized to see it, with zero error rate (nonzero error rates for security just aren’t acceptable). When valuable secure content isn’t included, the value that end users derive from the intranet search plummets, therefore so does the usage.

Happily, this is changing. When administrators can rely on error-free operations in terms of security, there is no reason, except for lack of budget, ambition, or imagination, to withhold the most valuable content from the intranet search.

Corporations like Organon, Modine Manufacturing, and others are seeing 10x-100x jumps in usage and value, right off the bat, from deploying cross-repository, unified search with real productivity increases for employees.

As an aside: I’d like to share a personal anecdote from my own intranet use. A prospective partner for Vivisimo recently came to my attention. Wishing to know what contact points we had had with this prospect, I searched for their company name on our cross-repository intranet search, which turned up a brief exchange dating back 1 1/2 years. I telephoned the individual (not a Vivisimo employee) involved in the exchange and got a valuable insight into the prospect, expending 15 total minutes. Absent our intranet search, my information need would have remained unsatisifed. In the course of a day, there are many questions like this which are left unanswered because people can’t fruitfully search their environment.

I’ll venture a prediction: daily enterprise searches per capita will surpass web searches by the end of 2010. With a reliable, unified intranet search in place, it’s hard not to envision knowledge workers doing many business-relevant searches per day in their own environments, posing heretofore unanswerable questions that are answered on the spot.

Search-based business intelligence for all, not just for the numerati.

Tags: business intelligence, desktop search, enterprise search, intranet search, web search

This entry was posted on Tuesday, February 19th, 2008 at 2:43 pm and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.